Marion Nicolaus

Schooling is a widespread behaviour among pelagic fish that reduces locomotion costs of migration and predation risk. It is characterised by coordinated movements, speed and orientation of group members. Although schooling can vary both between populations and among individuals of the same population, these levels of variation are seldom considered together. Here, we investigated whether schooling tendencies in three-spined sticklebacks ( Gasterosteus aculeatus ) vary between populations and among individuals within-populations depending on their phenotype. We first compared the schooling behaviour of anadromous (‘migrants’) and recently landlocked (‘residents’) populations and tested whether individual sociability explained within-population variation in schooling. We then studied whether schooling behaviour is modulated when migrants and residents are in mixed groups. Migrant groups displayed better schooling performance than residents with higher alignment. Within migrant groups, less social individuals displayed more leadership, were better aligned, swam closer to others and fatigued less than more social fish. These patterns were absent in residents. When the two populations were mixed, overall schooling performance declined, particularly in migrants, with schools exhibiting reduced compaction and alignment, resembling resident groups. Thus, individuals did not modulate their schooling behaviour to the group social composition and phenotype-dependent schooling patterns were disrupted in migrants when mixed. These findings suggest that migration-related behaviours can disappear quickly post-isolation. They also highlight the importance of individual sociability in self-organisation and shaping collective behaviour, especially in populations that retain migratory behaviour. Considering behavioural variation across multiple hierarchical levels is thus key to understanding complex social systems. Significance statement: Many animals exhibit complex social behaviours that can benefit group members. In fish, schooling behaviour enables to migrate with reduced costs of locomotion and predation risk. Yet, the benefits of schooling and schooling behaviour itself are shaped by the environmental conditions, and thus are strongly influenced by human disturbance. We conducted a study in three-spined sticklebacks to test whether man-made barriers that prohibit some fish populations to undergo their seasonal migration (‘residents’) altered schooling behaviour compared to their ‘migrant’ anadromous ancestors. Migrants were indeed better at schooling than residents, and their schooling patterns were shaped by individual variation in sociability. Residents did not improve their schooling behaviour when mixed with migrants. This suggests that ∼ 50 years of isolation are sufficient to losing migration-related behaviour and that river reconnection will likely promote the emergence of partial migration in this system.

Individual phenotypes strongly affect dispersal. Yet studying the relationship between dispersal tendencies and personality under ecologically relevant conditions is challenging, especially in fish. Laboratory studies lack environmental complexity and scale, while field-based approaches are often unfeasible. We here mimicked dispersal events in fish; encompassing all 3 phases of dispersal (departure, transience, settlement) in a large experimental mesocosm containing breeding sites. Using 3-spined sticklebacks (Gasterosteus aculeatus) tagged with ‘Passive Integrated Transponders’ (PIT tags), we remotely monitored proxies of dispersal of 2 cohorts, each consisting of 40 individuals of known activity and aggressiveness scores, continuously over 2 wk. We tested whether personality variation explained differences in departure timing, transience movements, settlement success and spatial distribution along the environmental gradient. We found that more active/aggressive individuals dispersed quicker in the mesocosm, spent more time on territories and collected more eggs. However, we did not detect strong phenotype-environment correlations. Our findings emphasize the potential role of personality-dependent dispersal in population dynamics, with specific behavioral types driving dispersal and settlement success. Our mesocosm setup provides unique insights into the dispersal dynamics in a system where detailed observations across all dispersal stages are rare and difficult to obtain. Despite limitations, behavioral experiments in mesocosms serve as a valuable bridge between laboratory and natural systems.

Migratory birds connect ecosystems worldwide and often show connectivity between breeding and nonbreeding areas, meaning that individuals have nonbreeding (“wintering”) locations close to others from the same breeding population. Although much work has characterized migration routes, we know little about what determines nonbreeding areas. We tracked pied flycatchers ( Ficedula hypoleuca ) from their entire breeding range and identified population-specific nonbreeding areas. Using an egg translocation experiment between the Netherlands and Sweden, we show that both inheritance and natal environment determine nonbreeding areas. Genetically Dutch birds that hatched in Sweden had nonbreeding areas intermediate between Swedish and Dutch populations. Although an inherited basis for migratory routes and nonbreeding areas may hamper rapid adaptation to a changing world, dispersing individuals can create new breeding–nonbreeding site combinations.

Fishways mitigate barriers blocking fish migration, but their effectiveness varies among species. Whether passage success also varies among individuals of the same species based on their morphology, physiology, or behaviour is largely unknown, particularly for small species. Yet, this knowledge is crucial for conservation practices due to potential artificial selection of pheno‐/genotypes with consequences for populations and ecosystems. This study tested whether a fishway induced a selective bias in exploration, length, and timing of migration in anadromous three‐spined sticklebacks ( Gasterosteus aculeatus ) in the Netherlands. Sticklebacks were PIT tagged, measured, tested for exploration, and released back near the fishway entrance. The fishway was monitored for 2 months. Passage success was 12.7% for all fish and 23% for those detected at the entrance, with fish passing 5.4 days after release on average. Passage success and time were unrelated to exploration or length. However, the timing of migration strongly predicted both. Individuals that were tagged later in the season passed the fishway faster and more successfully, likely due to increased reproductive demands and the need to access breeding grounds. Capture‐mark‐recapture analysis showed that increased passage was not explained by higher survival of late migrants, but by increased temperatures likely driving higher activity levels later in the season. High predation risk near the fishway may further explain the low overall passage success. Long‐term monitoring over multiple years, combined with prey–predator interaction studies, is needed to better evaluate the population‐level consequences of selective passage and predation pressure. This will help improve fishway design aimed at enhancing effectiveness.

Studying the relationship between dispersal tendencies and personality, or ‘dispersal syndromes’, under ecologically relevant conditions is challenging, especially in fish. Laboratory studies lack environmental complexity and scale, while field-based approaches are often unfeasible. We here mimicked dispersal events; encompassing all three phases of dispersal (departure, transience, settlement) in a large experimental mesocosm containing a gradient of low- and high-quality breeding sites. Using Passive Integrated Transponder (PIT)-tagged three-spined sticklebacks ( Gasterosteus aculeatus ), we remotely monitored the movements and behaviours of two groups, each consisting of 40 individuals of known personality, continuously over two weeks. We tested whether personality variation explained differences in departure timing, transience movements, settlement success and spatial distribution along the environmental gradient. We found evidence for a dispersal syndrome. Specifically, more active/aggressive individuals dispersed quicker in the mesocosm, spent more time on territories and collected more eggs. However, we did not detect phenotype-environment correlations. Our findings highlight the importance of personality-dependent dispersal, with specific behavioural types driving dispersal and settlement success. Furthermore, our mesocosm setup provides unique insights into dispersal dynamics in a system where detailed observations in all dispersal stages are rare. Despite limitations, behavioural experiments in mesocosms serve as a valuable bridge between laboratory and natural systems.

Habitat fragmentation can have negative impacts on migratory organisms that rely on the functional connectivity between growing and breeding grounds. Quantifying the population-level phenotypic consequences of such fragmentation requires fine-scaled tracking of individual behaviour and movements across relevant scales. Here we make use of a natural experiment where some populations of ‘migrant’ three-spined sticklebacks ( Gasterosteus aculeatus ) became ‘residents', following habitat fragmentation five decades ago. To test whether residents have a lower movement tendency than migrants, we developed a novel experimental platform that allows the automated tracking of individual movements via RFID technology in a semi-natural mesocosm where spatio-temporal scales and environmental conditions can be manipulated. We found that residents moved significantly less than migrants at large but not at small spatial scale. This pattern was consistent across time and contexts (water flow and group size). Our study substantiates prior literature on rapid phenotypic divergence in sticklebacks in response to human-induced isolation and highlights the importance of observing behaviour in ecologically relevant set-ups that bridge the gap between laboratory and field studies.

In most animal species, dispersing individuals possess phenotypic attributes that mitigate the costs of colonization and/or increase settlement success in new areas (dispersal syndromes). This phenotypic integration likely affects population dynamics and the direction of selection, but data are lacking for natural populations. Using an approach that combines population dynamics, quantitative genetics, and phenotypic selection analyses, we reveal the existence of dispersal syndromes in a pied flycatcher (Ficedula hypoleuca) population in the Netherlands: immigrants were larger, tended to have darker plumage, bred earlier, and produced larger clutches than local recruits, and some of these traits were genetically correlated. Over time, the phenotypic profile of the population gradually changed: each generation advanced arrival and breeding and exhibited longer wings as a result of direct and indirect selection on these correlated traits. Although phenotypic attributes of immigrants were favored by selection during the early phase of colonization, observed phenotypic changes were similar for immigrants and local recruits. We propose that immigrants facilitated initial population establishment but that temporal changes likely resulted from climate change-induced large-scale selection. This study highlights that newly established populations are of nonrandom composition and that phenotypic architecture affects evolutionary population trajectories.

In the 1970s, water management in the Netherlands resulted in numerous isolated populations of three-spined sticklebacks, which can no longer migrate from freshwater to the sea. We tested whether ∼50 years of isolation resulted in reduced migratory tendencies in these ‘resident’ sticklebacks. Lab-based individual testing showed behavioural divergence between residents and migrants, but also produced counter-intuitive results, especially with regards to movement tendencies. To detect differences in migration tendencies, we set up a semi-natural mesocosm, consisting of connected ponds, where movements of numerous individuals could continually be tracked at larger spatial scales. We found that wild-caught residents and migrants exhibited no differences in movement tendencies ‘within ponds’, but residents moved significantly less ‘between ponds’ than migrants. Between-pond movements were consistent and the observed differences were robust across contexts (changes in water flow and group size). Our study reveals that larger-scale movement tendencies can diverge over short time scales in response to human-induced isolation, and highlights the importance of observing behaviour in ecologically relevant setups that bridge the gap between lab and field studies.

Ecological research is often hampered by the inability to quantify animal diets. Diet composition can be tracked through DNA metabarcoding of fecal samples, but whether (complex) diets can be quantitatively determined with metabarcoding is still debated and needs validation using free‐living animals. This study validates that DNA metabarcoding of feces can retrieve actual ingested taxa, and most importantly, that read numbers retrieved from sequencing can also be used to quantify the relative biomass of dietary taxa. Validation was done with the hole‐nesting insectivorous Pied Flycatcher whose diet was quantified using camera footage. Size‐adjusted counts of food items delivered to nestlings were used as a proxy for provided biomass of prey orders and families, and subsequently, nestling feces were assessed through DNA metabarcoding. To explore potential effects of digestion, gizzard and lower intestine samples of freshly collected birds were subjected to DNA metabarcoding. For metabarcoding with Cytochrome Oxidase subunit I (COI), we modified published invertebrate COI primers LCO1490 and HCO1777, which reduced host reads to 0.03%, and amplified Arachnida DNA without significant changing the recovery of other arthropod taxa. DNA metabarcoding retrieved all commonly camera‐recorded taxa. Overall, and in each replicate year (N = 3), the relative scaled biomass of prey taxa and COI read numbers correlated at R = .85 (95CI:0.68–0.94) at order level and at R = .75 (CI:0.67–0.82) at family level. Similarity in arthropod community composition between gizzard and intestines suggested limited digestive bias. This DNA metabarcoding validation demonstrates that quantitative analyses of arthropod diet is possible. We discuss the ecological applications for insectivorous birds.